Archives of Biochemistry and Biophysics 569 (2015) 32–44

Contents lists available at ScienceDirect

Archives of Biochemistry and Biophysics

journal homepage: www.elsevier .com/ locate /yabbi

The invertebrate Caenorhabditis elegans biosynthesizes ascorbate

http://dx.doi.org/10.1016/j.abb.2015.02.0020003-9861/� 2015 Elsevier Inc. All rights reserved.

⇑ Corresponding author at: Department of Chemistry and Biochemistry and theMolecular Biology Institute, University of California, Los Angeles, 607 Charles E.Young Drive East, Los Angeles, CA, USA. Fax: +1 (310) 825 1968.

E-mail address: clarke@mbi.ucla.edu (S.G. Clarke).

1 Abbreviations used: HIF-1, hypoxia-inducible factor 1; HPLC, high-perfliquid chromatography; EIC, extracted ion current; TIC, total ion curreascorbic acid oxidase; ROS, reactive oxygen species; NAT, nucleobase atransporter.

Alexander N. Patananan, Lauren M. Budenholzer, Maria E. Pedraza, Eric R. Torres, Lital N. Adler,Steven G. Clarke ⇑Department of Chemistry and Biochemistry and the Molecular Biology Institute, UCLA, Los Angeles, CA 90095, USA

a r t i c l e i n f o a b s t r a c t

Article history:Received 7 December 2014and in revised form 28 January 2015Available online 7 February 2015

Keywords:Ascorbic acidVitamin CCaenorhabditis elegansBiosynthetic pathwayInvertebrates

L-Ascorbate, commonly known as vitamin C, serves as an antioxidant and cofactor essential for many bio-logical processes. Distinct ascorbate biosynthetic pathways have been established for animals and plants,but little is known about the presence or synthesis of this molecule in invertebrate species. We haveinvestigated ascorbate metabolism in the nematode Caenorhabditis elegans, where this molecule wouldbe expected to play roles in oxidative stress resistance and as cofactor in collagen and neurotransmittersynthesis. Using high-performance liquid chromatography and gas-chromatography mass spectrometry,we determined that ascorbate is present at low amounts in the egg stage, L1 larvae, and mixed animalpopulations, with the egg stage containing the highest concentrations. Incubating C. elegans with precur-sor molecules necessary for ascorbate synthesis in plants and animals did not significantly alter ascorbatelevels. Furthermore, bioinformatic analyses did not support the presence in C. elegans of either the plantor the animal biosynthetic pathway. However, we observed the complete 13C-labeling of ascorbate whenC. elegans was grown with 13C-labeled Escherichia coli as a food source. These results support the hypoth-esis that ascorbate biosynthesis in invertebrates may proceed by a novel pathway and lay the foundationfor a broader understanding of its biological role.

� 2015 Elsevier Inc. All rights reserved.

Introduction

L-Ascorbate (vitamin C) is a widely used metabolite in plantsand animals that functions as a cofactor and as an antioxidant byundergoing two successive one-electron oxidation reactions toyield monodehydroascorbate radical and dehydroascorbate [1,2].These reactions can participate in biosynthetic reactions, neutral-ize reactive oxygen species that lead to DNA, lipid, and proteindamage, and regulate processes involved in development andstress resistance [3–10]. Evidence has been presented that admin-istering increased amounts of ascorbate to humans can boost theimmune system [11], reduce the risk of chronic diseases [12,13],and improve the recovery of critically ill patients [14,15]. Inhumans and other vertebrates, ascorbate prevents the diseasescurvy by ensuring that the iron cofactors in the prolyl 3-, prolyl4-, and lysyl hydroxylases, enzymes responsible for the formationand stability of the collagen triple helix structure, are in the enzy-matically active, reduced ferrous state [16,17]. Other iron- and cop-per-dependent monooxygenases also rely on ascorbate as a

reductant, including the enzymes involved in the formation of neu-rotransmitters derived from tyrosine [17] and in carnitine biosyn-thesis [18,19], as well as the hypoxia-inducible factor 1 (HIF-1)1

that regulates metabolism in response to molecular oxygen levels[20].

The presence of ascorbate has been well documented in plantsand vertebrates. All photosynthetic organisms studied to date syn-thesize ascorbate [21], including higher plants, which can containup to 19% of their total soluble leaf carbon content as ascorbate[22,23], bryophytes [24], green algae [25], photosynthetic protists[22], and cyanobacteria [26,27]. In these organisms, the amountof ascorbate varies depending on the tissue analyzed, age of theorganism, time of day, altitude, and light intensity [22,28–30].Although ascorbate is present in cartilaginous, actinopterigian,and non-teleost fish [31,32], amphibians, and reptiles [33,34], itis not synthesized in all bat and bird species due to the inactivationof ascorbate biosynthesis genes [35]. Furthermore, humans,anthropoid primates, guinea pigs, and teleost fish cannot synthe-size ascorbate and require it in the diet [35].

ormancent; AAO,scorbate

A.N. Patananan et al. / Archives of Biochemistry and Biophysics 569 (2015) 32–44 33

In contrast to plants and animals, little is known about the pres-ence or the possible biosynthesis of ascorbate in invertebrates.Ascorbate was reported to be present in various marine inverte-brates [36–38] and in Drosophila melanogaster [39]. In Drosophila,ascorbate was detected in flies raised on an apparent ascorbate-free diet, but no evidence for endogenous biosynthesis of the mole-cule was shown [39]. Because no studies have recently exploredthe presence or possible biosynthesis of ascorbate in invertebrates,several authors have concluded that invertebrates are incapable ofsynthesizing this molecule [33,40,41]. Based on the importantroles of ascorbate in metabolism and in resistance to reactive oxy-gen species, we felt that it was important to establish in an inver-tebrate species both the levels of ascorbate and how it may bebiosynthesized.

Three distinct ascorbate biosynthetic pathways have now beendescribed for plants, vertebrates, and protist species. In plants, theSmirnoff–Wheeler pathway involves ten enzymes that participatein the conversion of D-glucose to L-ascorbate via GDP-D-mannoseand L-galactose [3]. VTC2, which catalyzes the conversion ofGDP-L-galactose to L-galactose-1-phosphate, regulates the firstcommitted step in this pathway [42]. In animals, ascorbate biosyn-thesis can be initiated with the conversion of UDP-glucuronate to

D-glucuronate [43]. After a reduction reaction that results in aninversion of configuration to give L-gulonate and a lactonizationreaction, ascorbate is formed by the L-gulono-1,4-lactone oxidase[43]. The absence of the gene encoding this oxidase in teleost fish,or its mutation in humans, anthropoid primates, guinea pigs, andother species, results in the loss of the ability to make ascorbatein these animals [35,44]. Interestingly, although some bat and birdspecies are unable to synthesize ascorbate as previously men-tioned due to the inactivation of this oxidase gene by mutation,other species have incurred further mutations that have reactivat-ed the oxidase gene and restored ascorbate synthesis [35]. A thirdbiosynthetic pathway using D-galacturonate has been reported insome protist and green plant species [21,45–47]. Finally, it hasbeen determined that fungi do not synthesize ascorbate, but ratherproduce D-erythroascorbate, a molecule with similar antioxidantproperties [48–51]. The pathway by which this molecule is madehas not been fully elucidated, but it is believed that the formationof D-erythroascorbate proceeds through D-arabinose [43,48].

Although there has been abundant research on the functions ofascorbate, it was not until recently that the plant, animal, and pro-tist biosynthesis pathways were clarified [3,21,43]. Nevertheless,modifications to these pathways are still being made to accommo-date new intermediates and interactions between the pathways[28,52], opening the possibility that other ascorbate biosynthesispathways may exist in other organisms.

In the present work, we focused on ascorbate in the inverte-brate nematode worm Caenorhabditis elegans. It was previouslydiscovered by serendipity that the C10F3.4 (mcp-1) gene is a homo-log to the plant VTC2 gene that encodes the enzyme catalyzing therate-limiting step in the plant ascorbate biosynthetic pathway[42,53]. This observation prompted the question of whether ornot ascorbate is present in C. elegans. Although the mcp-1 genehas a function in C. elegans distinct from ascorbate biosynthesis[53], ascorbate was suggested to be present in worm extracts[42,53,54]. In the present work, we expand upon these findingsusing high-performance liquid chromatography and gas chro-matography–mass spectrometry approaches to quantitate ascor-bate in C. elegans. We now show that ascorbate is present andsynthesized by C. elegans, thereby providing evidence that inverte-brate organisms are capable of synthesizing this importantmolecule.

Materials and methods

C. elegans husbandry

Bristol N2 C. elegans were used in this study and obtained fromthe Caenorhabditis Genetics Center (University of Minnesota, SaintPaul, MN). C10F3.4 (mcp-1) mutant strain tm2679 animals wereobtained from the C. elegans Core Facility at Tokyo Women’s Med-ical University. The strains were maintained at 20 or 25 �C on10 cm � 1.5 cm petri dish plates containing nematode growthmedium (NGM, 20 g/l bacto agar (BD Biosciences, catalog #214030), 2.5 g/l bacto peptone (BD Biosciences, catalog #211820), 3 g/l NaCl, 1 ml of 1 M CaCl2 per liter, 1 ml of 1 M MgSO4

per liter, 25 ml of 1 M potassium phosphate, pH 6, per liter, and1 ml of 5 mg/ml cholesterol (prepared in ethanol) per liter). Eachplate contained a spot of about 150 ll of streptomycin-resistantEscherichia coli OP50-1 (approximately 5 � 1010 cells/ml) grownin Luria Broth (LB, BD Biosciences, catalog # 244610) supplement-ed with 50 lg/ml streptomycin (Sigma, catalog # S6501). Theworms on these plates were fed every 3–4 days with OP50-1 andmaintained for 2–3 weeks.

Egg harvesting by bleach treatment

Mixed populations of N2 animals were transferred from main-tenance plates to approximately 100 fresh NGM plates containingOP50 E. coli and were incubated at 25 �C for 3 days to obtain gravidadults. The plates were washed with ice-cold M9 minimal medium(22 mM KH2PO4, 42 mM Na2HPO4, 86 mM NaCl, and 1 mM MgSO4)and the animals were transferred to 15 ml polypropylene conicaltubes. The worms were centrifuged at 600�g for 2 min at 4 �C(Beckman Coulter, Allegra X-15R) and the supernatant was aspirat-ed. The animals were resuspended in 10 ml of ice-cold 30% sucroseand centrifuged at 2095�g for 5 min at 4 �C. The floated animalswere transferred to a 15 ml conical tube with ice-cold M9 medium,pelleted at 2000�g for 2 min, and the supernatant was aspirated.The worms were subsequently washed approximately 5 times withice-cold M9 medium and by centrifuging at 600�g for 2 min. Afterthe final wash, the worm pellet was resuspended in 5 ml of freshbleaching solution (12.5 ml Clorox bleach (6.2 percent sodiumhypochlorite), 6 ml 5 M NaOH, 31.5 ml water) and pelleted in aclinical centrifuge at setting 5 (�796�g) for 30 s (InternationalClinical Centrifuge, model CL). The supernatant was aspiratedand the pellet resuspended in 10 ml of bleaching solution. Theworms were fragmented by alternating between 30 s of vortexingat setting 10 and 15 s on ice, and the release of eggs from gravidadults was monitored by microscopy. After no more than 6 min,the samples were centrifuged in a clinical centrifuge at setting 5for 30 s and the eggs in the pellet were washed with ice-cold M9medium 5 times to remove residual bleaching solution. C. eleganseggs were analyzed directly for ascorbate content or used to gener-ate L1 larvae as described below.

Sample preparation for high-performance liquid chromatography(HPLC) ascorbate analysis

Mixed animal populations were washed from 2 to 80 NGMplates with ice-cold M9 medium into a 15 ml conical tube and cen-trifuged at 600�g for 2 min. The supernatant was aspirated and theanimals were washed an additional time with M9 medium. Thepellet was transferred to a microcentrifuge tube, weighed, andresuspended in 150 ll of ascorbate extraction buffer consisting of5% meta-phosphoric acid (Sigma, catalog # M6288), 2 mM disodi-um EDTA, and 2 mM tris(2-carboxyethyl) phosphine hydrochloride(Sigma, catalog # C4706). This reagent reduces dehydroascorbate

34 A.N. Patananan et al. / Archives of Biochemistry and Biophysics 569 (2015) 32–44

to ascorbate; our assays then reflect the amount of both of thesemolecules. Approximately 50–100 ll of washed sea sand (FisherScientific, catalog # S25) was added and the samples were groundon ice with a disposable pestle (Fisher Scientific, catalog # 12–141-364) for 100 revolutions by hand or using a lab stirrer for 1 min setto 100 rpm (Fisher Scientific, LR400A). The sand was removed andthe samples were centrifuged at 20,000�g for 10 min at 4 �C(Eppendorf, Model 5417R). The supernatant was transferred to anew microcentrifuge tube and stored for no more than 24 h at�80 �C until analysis to minimize sample degradation. Ascorbatelevels in eggs were determined from approximately 80 NGM platesby performing the bleach treatment and lysis procedure describedabove. Eggs were allowed to hatch in 70 ml of M9 medium at 20 �Cand 160 rpm for at least 24 h to determine the presence of ascor-bate in L1 larvae.

In general, 60 ll of extract or L-ascorbate standard (Fluka, cata-log # 95209) dissolved in ascorbate extraction buffer were adjust-ed to a final pH of �5 by adding 9 ll of 1 M sodium citrate, �pH 8(final concentration of 130 mM), and separated on a HP 1090 II liq-uid chromatograph with a Phenomenex C18 reversed-phase col-umn (Catalog # 00G-4252-E0, 5 lm, 4.6-mm inner diameter,250-mm length). Analyses were performed at room temperaturewith an injection volume of 30–55 ll and a flow rate of 1 ml/min. Buffer A consisted of 20 mM triethylammonium acetate inwater, pH 6, and buffer B was 20 mM triethylammonium acetatein 40% acetonitrile, pH 6. The HPLC gradient was as follows: iso-cratic for 7 min at 100% A, a 1 min linear gradient from 100% Ato 0% A, 5 min at 0% A, a 1 min linear gradient from 0% A to 100%A, and isocratic for 15 min at 100% A. Ascorbate was detected ata wavelength of 265 nm and eluted between 3.6 and 4 mindepending on the column condition. The presence of ascorbatewas verified by noting the loss of the peak after 60 ll of samplewas incubated with 2 units of ascorbate oxidase (AAO, Fisher Sci-entific, catalog # 50-230-3657, dissolved in 100 mM sodium phos-phate and 0.5 mM EDTA, pH 5.6) and 9 ll of 1 M sodium citrate, pH8, for 30–120 min and chromatographed. The amount of ascorbatewas calculated based on the area of the ascorbate peak at 265 nmusing an extinction coefficient of 14,500 M�1 cm�1. Ascorbatelevels in samples were normalized either to the approximate num-ber of worms analyzed as determined by microscopy, or by assum-ing 1 g of wet weight worm pellet was equivalent to 1 ml of wormvolume.

Addition of ascorbate precursors to C. elegans extracts and intactanimals

For incubations with gulono-1,4-lactone with fractions ofworms, mixed populations of animals were washed with ice-coldM9 medium from 20 NGM plates and pelleted in a 15 ml conicaltube at 600�g and 4 �C for 2 min. After the supernatant was aspi-rated, the pellet was washed two times with ice-cold M9 medium.The pellet was weighed (�410 mg) and 820 ll of protein extractionbuffer (10 mM Tris–HCl, pH 7.4, 1.5 mM MgCl2, 10 mM KCl, 50 mMsucrose, 1 mM DTT) was added. The animals were freeze thawedthree times with liquid nitrogen before lysis (Branson Sonic PowerCo., W-350 Sonifier) on ice with 10 5-s sonicator pulses (50% dutycycle, intensity 4) in a total time of 5 min. The samples were placedin microcentrifuge tubes and centrifuged at 600�g for 1 min toobtain the ‘‘intact worm’’ pellet of unbroken and fragmented ani-mals. The supernatant was transferred to new microcentrifugetubes and centrifuged at 20,000�g for 15 min to obtain the ‘‘wormmembrane’’ pellet. The supernatant was transferred to a newmicrocentrifuge tube and labeled ‘‘worm cytosol.’’ An additional100 ll of protein extraction buffer was added to the ‘‘intact worm’’and ‘‘worm membrane’’ pellets. In a total volume of 75 ll, 25 ll of‘‘intact worms,’’ ‘‘worm membrane,’’ or ‘‘worm cytosol’’ were incu-

bated in 20 ll of water, 5 ll of 100 mM L-gulono-1,4-lactone (Sig-ma, catalog # 310301, final concentration of 6.7 mM), and 25 ll of100 mM sodium phosphate and 0.5 mM EDTA, pH 5.6, for 2 h atroom temperature. The reactions were terminated by adding75 ll of 10% trifluoroacetic acid and 140 ll of sample was injectedon the HPLC and ascorbate was characterized as described above.To confirm the position of ascorbate, 2 units of AAO was incubatedwith the above reactions for 1 h at 4 �C and separated by HPLC.

For incubations of live worms with gulono-1,4-lactone andother possible ascorbate precursors, eggs were obtained from 80NGM plates containing a mixed population of animals as describedabove and incubated in 5 ml of M9 medium at 20 �C and 160 rpmfor approximately 24 h. The L1 larvae were centrifuged at 3000�gfor 5 min and the supernatant was aspirated. The worms werewashed once with ice-cold M9 medium and a pellet with L1 larvaeand residual M9 medium (�900 ll) was obtained. 100 ll of L1 lar-vae in suspension were subsequently incubated with 20 ll ofeither protein extraction buffer, 10 mM L-gulono-1,4-lactone,10 mM D-(+)-galacturonic acid (Fluka, catalog # 48280), 10 mM

D-glucuronic acid (Sigma, catalog # G5269), 10 mM D-galactose(Fisher Scientific, catalog # BP656), or 10 mM D-glucose (Fisher Sci-entific, catalog # D16). After a 2 h incubation at 20 �C and 160 rpm,50 ll of washed sea sand was added to each reaction to lyse theanimals and the ascorbate content was analyzed by HPLC asdescribed above.

Gas chromatography–mass spectrometry (GC–MS) analysis

After obtaining the animals as described above, pellets ofworms were transferred to a microcentrifuge tubes and 150–300 ll of GC–MS extraction buffer (0.5 mM butylated hydroxyto-luene (BHT, Sigma, catalog # B1378) in methanol) was added.Approximately 50–100 ll of washed sea sand was added to eachsample and the animals were lysed with 100 rotations of a plasticpestle. After the sand was removed, the samples were centrifugedat 20,000�g for 10 min. The supernatant was removed and trans-ferred to amber microcentrifuge tubes until use.

Acid washed, amber micro-v glass vials (Thermo Scientific, cat-alog # C4000-V2) were used for the GC–MS analysis. Approximate-ly 300–800 ll of sample was dried in GC vials in an unheatedvacuum centrifuge (Savant Instruments Inc., model # SVC100Hand RH 20-12 rotor). After the samples were dried, 100 ll of ben-zene was added to the vials and the samples were dried undernitrogen gas for approximately 15 min to remove residual water.The samples were derivatized in 50 ll of a 99:1 mixture of N,O-bis(trimethylsilyl)trifluoracetamide:trimethylchlorosilane (BSTFA/TMCS) (Supelco, catalog # 33148) at 60 �C for 90 min. Ascorbatestandards were prepared in 0.5 mM BHT in methanol and deriva-tized using the same protocol described above.

Samples were analyzed using an Agilent 6890 GC–MS with5975 mass selective detector and helium carrier gas. The autosam-pler syringe was washed 3 times with chloroform before and aftereach injection. 1 ll of derivatized sample was injected and the GC–MS was set to have a 50:1 split injection, helium flow rate of1.3 ml/min, and a 250 �C injection port temperature. Samples wereseparated on an Agilent JW HP-5-MS column (30 m length, 250 lminner diameter, and 0.25 lm film thickness) with the following17.7 min method: 1 min at 140 �C, increase temperature 15 �C/min until 300 �C, 6 min at 300 �C. A 3 min solvent delay wasapplied to prolong detector life. GC–MS chromatograms were ana-lyzed and molecules identified using the Agilent Enhanced Chem-station, NIST Mass Spectral Search Program 2.0f, and the NIST MassSpectral Library 08. Ascorbate eluted at approximately 8.6–8.7 minand levels in C. elegans samples were normalized by comparing the

Fig. 1. HPLC detection of ascorbate in extracts of mixed populations of C. elegans.(A) Ascorbate standard separated by reversed-phase HPLC. The black line represents3.9 lg of ascorbic acid dissolved in a final concentration of 1.7% metaphosphoricacid, 1.7 mM EDTA, 1.7 mM tris(2-carboxyethyl) phosphine hydrochloride, and130 mM sodium citrate, pH 8, and separated on a reversed-phase HPLC column asdescribed in the ‘‘Materials and methods’’ section. 2 units of ascorbate oxidase(AAO) were added to the standard to confirm the elution of ascorbate as describedin ‘‘Materials and methods’’ (red line). (B) Reversed-phase HPLC analysis of a C.elegans extract prepared from a mixed population of animals obtained by washingtwo NGM plates (�142,000 worms) as described in the ‘‘Materials and methods.’’55 ll of sample was injected onto the HPLC (black line). 2 units of AAO was added toconfirm the elution of ascorbate (red line). (For interpretation of the references tocolor in this figure legend, the reader is referred to the web version of this article.)

A.N. Patananan et al. / Archives of Biochemistry and Biophysics 569 (2015) 32–44 35

ratio between the extracted ion current (EIC) areas of ascorbate(332 m/z) to serine (204 m/z).

To determine if ascorbate is present in the buffers and E. coliused to grow C. elegans, 100 ll of LB, 400 ll of E. coli (�8 � 109 -cells) lysed by sonication in 0.5 mM BHT in methanol, 100 ll of a5% sucrose solution in water, 200 ll of M9 medium, 200 ll of Smedium (to every 100 ml of S Basal (5.84 g/l NaCl, 50 ml of 1 MKH2PO4, pH 6, per liter, 1 ml of 5 mg/ml cholesterol (prepared inethanol) per liter) the following components were added: 1 ml of1 M potassium citrate, pH 6, 1 ml of trace metals solution, 0.3 mlof 1 M CaCl2, and 0.3 ml of 1 M MgS04), 100 ll of 0.5 mM BHT inmethanol, and a mixed population of C. elegans extract washedfrom 10 NGM plates and prepared as previously described, weredried by vacuum centrifugation and derivatized as describedabove. The total ion current (TIC) and EIC of these samples wereanalyzed by GC–MS.

Paraquat and ethanol stress on C. elegans ascorbate levels

A mixed population of animals were collected from approximate-ly 6 NGM plates, washed without a sucrose float, and incubated inflasks with 100 mL of S medium and OP50 E. coli at 20 �C and160 rpm. After approximately 24 h, paraquat (Sigma, catalog #856177) was added and the flasks were allowed to incubate foranother 7 days. The samples were subsequently centrifuged at600�g for 2 min, the supernatant was aspirated, and the pellet waswashed once with M9 medium. 0.5 mM BHT in methanol was addedto the pellet in a 1:1 ratio (g pellet:ml extraction buffer) and lysed bysonication as described above. GC–MS analysis was performed asdescribed above using 400 ll of extract. For ethanol stress, a mixedpopulation of animals was obtained from 80 NGM plates asdescribed above. Approximately 250 ll of worms were transferredto microcentrifuge tubes with ethanol (Acros, 200 proof, catalog #61509-0010) diluted in M9 medium. After a 39 h incubation, the ani-mals were pelleted, washed, lysed, and analyzed for ascorbate byreversed-phase HPLC as described above.

Labeling of C. elegans using 13C-labeled OP50 E. coli

5 ml of LB medium supplemented with 50 lg/ml of strepto-mycin was inoculated with OP50 E. coli. After an incubation at37 �C and 250 rpm for approximately 8 h, 100 ll of the culturewas transferred to a flask containing 12.5 ml of a E. coli M9 mini-mal medium with 10.9 mM D-glucose-13C6 (Cambridge IsotopeLaboratories, catalog # CLM-1396–0), 93.7 mM KH2PO4, 56.3 mMK2HPO4, 62.2 mM Na2HPO4, 13.5 mM K2SO4, 20.6 mM NH4Cl,17.5 mM uracil, 4.9 mM MgCl2, 0.2 mM CaCl2, 0.1 mM FeSO4,28.5 lM MnCl2, 16.5 lM CoCl2, 11.9 lM ZnSO4, 8.6 lM CuCl2,1.6 lM H3BO3, 4.6 lM (NH4)6Mo7O24, and 65.8 lM EDTA. After thisculture was incubated at 37 �C and 250 rpm for approximately14 h, all 12.5 ml were transferred to a flask containing 497.5 mlof the 13C-glucose containing medium described above and cul-tured for 9 h. The cell pellet (�1.5 g wet weight) was obtained bycentrifugation at 6000�g for 5 min at 4 �C and washed once withC. elegans M9 medium before it was resuspended in 1.5 ml C. ele-gans M9 medium. 150 ll of this concentrated E. coli was spottedonto NGM agarose plates and allowed to dry. For animals incubat-ed in liquid culture, 1.5 ml of concentrated E. coli was added to70 ml of M9 medium in a 250 ml beveled flask. Unlabeled E. colicells were prepared using the same protocol with the exceptionthat D-glucose (Fisher Scientific, catalog # D16) was substitutedfor the 13C-labeled glucose.

Mixed populations of C. elegans were used in the labelingexperiments and prepared as described above from 46 NGM plates.100 ll of animals (�1200 mixed stage worms and �2970 eggs)were pipetted onto each of 10 NGM agarose plates and allowed

to incubate in a 20 �C incubator for 85 h. For liquid culture, 1 mlof animals (�12,000 mixed stage worms and �29,700 eggs) wereadded to the 70 ml flasks and incubated at 20 �C and 160 rpm for85 h. The animals were harvested, washed, and analyzed by GC–MS as described above.

Results

Ascorbate is detected in N2 C. elegans extracts and is not present ingrowth media, E. coli cells, or the extraction reagent

HPLC and LC-MS/MS had previously been used to provide evi-dence for the existence of ascorbate in C. elegans [42,53,54]. In thiswork, we optimized a reversed-phase HPLC method to detectascorbate, which eluted on our system between 3.6 and 4 min(Fig. 1A). To verify the presence of ascorbate, ascorbate oxidase(AAO) was used to convert ascorbate into dehydroascorbate, whichhas a decreased absorbance at 265 nm [55,56] (Fig. 1A). When anextract from a mixed population of N2 animals was analyzed, aspecies was detected at an elution time similar to ascorbate(Fig. 1B). The addition of AAO to this sample resulted in adecreased absorbance at 265 nm, suggesting that this peak wasascorbate (Fig. 1B). To verify the presence of ascorbate, we

Fig. 2. GC–MS detection of ascorbate in mixed populations of C. elegans. (A) GC–MS fragmentation pattern of ascorbate derivatized with BSTFA/TMCS from the NIST MassSpectral Library 08. An enlarged view of the m/z for the specific fragments labeled a–g are shown in the panels at the top. In the structures represented below the panel, thefragment of ascorbate is shown in red; when multiple fragments are possible for a particular m/z, the alternative fragment is shown in blue. Among the peaks, the 332 m/zfragment represents one of the most intense signals. (B). GC–MS analysis of a BSTFA/TMCS derivatized extract of mixed population of C. elegans obtained from ten NGM plateswas performed as described in the ‘‘Materials and methods.’’ The total ion current (TIC) and 332 m/z extracted ion current [97] are represented by the black and red traces,respectively. The asterisk (*) denotes the ascorbate peak. (C) The MS fragmentation pattern of the species denoted by the asterisk identified in panel B. (D) GC–MS analysis of astandard of 150 ng of ascorbate derivatized in BSTFA/TMCS to confirm the elution time of ascorbate as described in the ‘‘Materials and methods.’’ The TIC and 332 m/z EIC arerepresented by the black and red traces, respectively. The asterisk (*) denotes the ascorbate peak. The peak located at 7.4 min is an artifact identified as derivatized triethyleneglycol. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

36 A.N. Patananan et al. / Archives of Biochemistry and Biophysics 569 (2015) 32–44

analyzed extracts of C. elegans by GC–MS. Ascorbate derivatized inBSTFA/TMCS displays several major GC–MS fragment ions fromdata in the NIST mass spectral library, including a relatively abun-dant species with a peak at 332 m/z (Fig. 2A). We therefore ana-lyzed a mixed population of C. elegans by GC–MS and found astrong 332 m/z EIC signal at 8.7 min (Fig. 2B). An analysis of thefragmentation pattern of the material eluting at 8.7 min (Fig. 2C)revealed a pattern matched to the ascorbate standard found inthe NIST mass spectral library with a probability score of 93.9%, aMatch score of 799 (comparison of the sample to the library spec-trum; 700–800, fair match; >800, good match), and an R. Matchscore of 848 (comparison that ignores peaks in the sample not pre-sent in the library spectrum) (Fig. 2C). To verify that ascorbateelutes at this position, we analyzed a derivatized ascorbate stan-dard by GC–MS and found a species eluting at 8.7 min identifiedin the NIST library as ascorbate with high probability (94.9%, Matchscore of 740, R. Match score of 747) (Fig. 2D). We found no evi-dence for the presence in C. elegans of D-isoascorbate, the 5-epimer

of L-ascorbate, which can be resolved from ascorbate on the GC sys-tem (data not shown). Combined, these results demonstrate thepresence of ascorbate in C. elegans extracts.

It is possible that some or all of the observed ascorbate wasderived from the extraction solution or the C. elegans growth mediacomponents. We thus investigated if the OP50 bacterial foodsource of C. elegans or the LB medium used to grow the bacteriacontains ascorbate. In addition, we also analyzed C. elegans growthmedia (M9 and S media), the sucrose involved in separating ani-mals from E. coli, and the extraction buffer (0.5 mM BHT in metha-nol) to determine if ascorbate was present. However, GC–MSanalysis did not indicate the presence of ascorbate in any of thesolutions or E. coli extracts tested (Fig. 3A and B). Although sucroseand E. coli extracts did have a small EIC signal at the 332 m/z, thefragmentation pattern at this elution time did not match ascorbate(Fig. 3C and D). Taken together, these HPLC and GC–MS analyseshave confirmed that C. elegans contain ascorbate and that it isnot a derived from the media or the bacterial food source.

Fig. 3. Lack of detectable ascorbate in C. elegans growth media and extraction reagent. (A) TIC and (B) 332 m/z EIC were analyzed by GC–MS of an extract derived from a mixedpopulation of C. elegans, 0.5 mM BHT in methanol extraction buffer, S medium, M9 medium, sucrose, E. coli extract, and LB medium. Samples were derivatized with BSTFA/TMCS as described in the ‘‘Materials and methods.’’ The dashed red line denotes the elution position of ascorbate. Although small peaks were found in the TIC and 332 m/z EICof sucrose and E. coli, further analysis of the mass spectrum at these elution times for the sucrose (C) and E. coli (D) samples reveal that the fragmentation patterns do notmatch ascorbate, but rather weakly match derivatized ribofuranose and phosphate, respectively. (For interpretation of the references to color in this figure legend, the readeris referred to the web version of this article.)

A.N. Patananan et al. / Archives of Biochemistry and Biophysics 569 (2015) 32–44 37

mcp-1 (tm2679) animals have similar ascorbate levels to wild type N2animals

MCP-1 (C10F3.4) is a C. elegans protein with homology to VTC2,an enzyme in the ascorbate biosynthesis pathway in plants[3,42,53,57,58]. Because plants deficient in VTC2 have decreasedascorbate levels [58], we hypothesized that if MCP-1 was impor-tant for ascorbate biosynthesis in C. elegans, mutants should alsoresult in decreased levels. We therefore tested the ascorbate levelsin mcp-1 (tm2679) deletion animals and compared them to thosefound in wild type. The analysis of mcp-1 (tm2679) extracts byHPLC confirmed the presence of ascorbate in animals (Fig. 4A)and the levels were quantified to be 15 ± 7 fmol of ascor-bate/worm, a concentration not significantly different to N2 ani-mals (36 ± 16 fmol/worm) (Fig. 4B). To further confirm ourfindings, we used GC–MS to determine the ratio of ascorbate(332 m/z) to serine (204 m/z) in samples. Our analysis did not finda statistically significant difference between the ratio in N2(0.024 ± 0.008) and mcp-1 (tm2679) (0.045 ± 0.04) animals. Theseresults suggest MCP-1 is not essential for C. elegans ascorbatebiosynthesis. A similar conclusion was previously reached withother methods [53,54].

Ascorbate levels vary between growth stages

Previous research has indicated that ascorbate levels declinewith age in organisms, including humans [59], rats [60], plants

[29], and potentially Drosophila [39]. To determine if levels alsodecline in C. elegans, we analyzed ascorbate concentrations at dif-ferent growth stages. The HPLC analysis of C. elegans extractsderived from eggs, L1 larvae, and mixed populations found ascor-bate levels to be the highest in the egg stage (0.144 ± 0.023 mM),compared to L1 larvae (0.055 ± 0.028 mM) and mixed population(0.021 ± 0.014 mM) (Fig. 5). These levels of ascorbate are low com-pared to higher plants and vertebrates, but are only somewhatlower than those found in the algae Chlamydomonas reinhardtii(Table 1). Ascorbate levels in C. elegans are also comparable tothose of erythroascorbate in the yeast Saccharomyces cerevisiae(Table 1).

Sugar precursors from known ascorbate biosynthetic pathways andoxidative stressors do not increase ascorbate levels

Plants and animals utilize one of two major alternative ascor-bate biosynthetic pathways. To determine if there are proteins inC. elegans that are homologous to those associated with ascorbatesynthesis in plants or animals, we compared C. elegans by proteinBLAST search to Mus musculus and Arabidopsis thaliana. In the com-parison to M. musculus, three proteins, UDP-glucose pyrophospho-rylase, UDP-glucose dehydrogenase, and b-glucuronidase,represented apparent orthologs with mutual best hits (Table 2).Additionally, the mouse UDP-glucuronosyltransferase andglucuronate reductase appear to have similar species in wormsalthough they are not mutual best hits. Most importantly,

Fig. 4. C. elegans deficient in mcp-1 have ascorbate levels equivalent to wild type.(A) mcp-1 (tm2679) extracts were prepared from two NGM plates containing amixed population of worms (approximately 208,000 animals) and analyzed byreversed-phase HPLC as described in the ‘‘Materials and methods’’ (black line). Theextracts were treated with 2 units of AAO to confirm the existence of ascorbate (redline). (B) Ascorbate levels were quantified by HPLC and compared between the N2wild type and mcp-1 (tm2679) extracts prepared from two NGM plates containing amixed population of worms (N2, 43,700–142,400 animals; mcp-1 (tm2679),106,000–318,500) as described in the ‘‘Materials and methods.’’ The pointsrepresent three independent samples, the horizontal line the average, and theerror bars represent standard deviations. Statistical significance was determinedusing an unpaired, two-tailed Student’s t-test. (C) N2 and mcp-1 (tm2679) extractswere prepared in 0.5 mM BHT in methanol and analyzed by GC–MS as described inthe ‘‘Materials and methods.’’ The area of the ascorbate 332 m/z EIC peak in eachsample was determined and compared against the area of the serine 204 m/z EICpeak. The points represent four independent samples, the horizontal line is theaverage, and the error bars represent standard deviations. Statistical significancewas determined using an unpaired, two-tailed Student’s t-test. (For interpretationof the references to color in this figure legend, the reader is referred to the webversion of this article.)

38 A.N. Patananan et al. / Archives of Biochemistry and Biophysics 569 (2015) 32–44

however, the final two enzymes in the animal biosynthetic path-way (the lactonase and oxidase) do not appear to be found inworms (Table 2). A comparison to the plant biosynthetic pathwaylikewise identified probable C. elegans enzymes for eight steps ofthe pathway. However, there was no apparent ortholog in wormsof the plant epimerase and the final L-galactono-1,4-lactone dehy-drogenase. These results suggest that a complete set of genesencoding all of the enzymes of either the animal or plant pathwayis not present in C. elegans. Finally, D-galacturonate has been deter-mined to be an intermediate in a third pathway of ascorbatebiosynthesis in some protists and green plants in a reaction form-ing L-galactonate [21,45–47]. Although C. elegans has an apparentortholog of the strawberry enzyme catalyzing this step(Y39G8B.1), it still appears to lack the final enzyme that is sharedwith the plant pathway to form ascorbate.

We wanted to explore the possibility that C. elegans may stillhave the missing enzyme activities of either the animal or plantpathway expressed by more distant or even unrelated genes. Todirectly address this question, mixed populations of animals werefirst prepared either as intact animals, or as membrane and cytoso-lic protein fractions from cell lysates. These materials were subse-quently incubated with the last precursor of the animal pathway, L-gulono-1,4-lactone, and analyzed by HPLC. In comparison to themembrane fraction alone, membranes incubated in the presenceof the sugar did not have significantly altered ascorbate levelsunder our assay conditions (Fig. 6A). A similar result was observedfor the cytosolic protein fraction and intact animals (Fig. 6B and C).We next tested if live animals could use sugars from both the plantand animal pathways to synthesize ascorbate. In order to analyze auniform population and eliminate the possibility that the bacteriapresent as a food source would use the ascorbate precursors,starved L1 larvae that were prepared in media lacking bacteriawere incubated with D-galactose, L-gulono-1,4-lactone, D-galactur-onic acid, D-glucose, or D-glucuronic acid. Once again, no differencein the level of ascorbate was observed by HPLC with the addition ofthe sugar precursors in comparison to the control (Fig. 6D).Although it is possible that our in vitro incubation conditionsand/or the gene expression status of the live worms were notconducive to detect ascorbate formed from the added precursors,

Fig. 5. Ascorbate levels are highest in the egg stage of C. elegans. Extracts of eggs, L1larvae, and mixed populations derived from 9–83, 5–69, and 38–455 mg wet weightof animals, respectively, were prepared as described in the ‘‘Materials and methods’’section. Ascorbate was separated by reversed-phase HPLC and quantified. Five,sixteen, and twelve independent samples were analyzed for the eggs, L1 larvae, andmixed population of worms, respectively. The data points represent independentsamples, the horizontal line is the average, and the error bars represent standarddeviations. Statistical significance was determined using an unpaired, two-tailedStudent’s t-test.

Table 1Comparison of ascorbate levels in C. elegans to other organisms.

Organism Ascorbate (mM)a References

Myrciaria dubia (camu camu) 136–170 Justi et al. [98]Gest et al. [22]

Arabidopsis thaliana (flowering plant) 4–9 Smirnoff et al. [21]Terrapene carolina (box turtle) 1.6–5.5 Rice et al. [34]Thamnophis sirtalis (common garter snake) 1.1–3.7 Rice et al. [34]Fragaria ananassa (strawberry) 1.5–3.3 Tulipani et al. [99]

Wang et al. [100]Gest et al. [22]

Mus musculus (mouse) 0.4–3 Tsao et al. [101]Xenopus laevis (African clawed frog) 1–2.3 Rice et al. [34]Himantura signifer (freshwater stingray) 0.04–1.2 Wong et al. [41]Chlamydomonas reinhardtii (green algae) 0.1–0.8 Urzica et al. [25]Saccharomyces cerevisiaeb (budding yeast) <0.1–0.4 Spickett et al. [102]Caenorhabditis elegans (nematode) 0.144 ± 0.023 (eggs)

0.055 ± 0.028 (L1 larvae)0.021 ± 0.014 (Mixed population)

This study

a Ascorbate concentrations from the literature were converted to mmol/l by assuming 1 gram of fresh/wet weight material is equivalentto 1 ml.

b Concentration shown is that of the ascorbic acid analog erythroascorbic acid.

Table 2Comparison of the animal and plant ascorbate pathways to C. elegans.

Protein UniprotID

Reaction C. elegansProtein

% AAidentity

E-value Mutual besthit

Animal (Mus musculus)UDP-glucose pyrophosphorylaseb Q91ZJ5 D-glucose-1-phosphate ? UDP-glucose K08E3.5d 67 0 Yes

UDP-glucose dehydrogenasea O70475 UDP-glucose ? UDP-glucuronate F29F11.1a 66 0 YesUDP-glucuronosyltransferasea Q62452 UDP-glucuronate ? acceptor-beta-D-

glucuronosideC08F11.8c 27 7x10�56 2nd

b-Glucuronidaseb P12265 a beta-D-glucuronoside ? D-glucuronate Y105E8B.9d 40 9x10�154 YesGlucuronate reductaseb Q540D7 D-glucuronate ? L-gulonate Y39G8B.1d 47 4x10�97 2nd

Gulonolactonasea Q64374 L-gulonate ? L-gulono-1,4-lactone EO3H4.3d 28 1.2 51st

L-gulonolactone oxidasea P58710 L-gulono-1,4-lactone ? L-ascorbate F45D5.12d 27 7x10�9 25th

Plant (Arabidopsis thaliana)Hexokinasea Q42525 D-glucose ? D-glucose-6-P F14B4.2a 37 1x10�79 Yes

Phosphoglucose isomerasea P34795 D-glucose-6-P ? D-fructose-6-P Y87G2A.8b 47 1x10�151 Yes

Phosphomannose isomerasea Q9M884 D-fructose-6-P ? D-mannose-6-P ZK632.4c 34 6x10�71 Yes

Phosphomannomutasea O80840 D-mannose-6-P ? D-mannose-1-P F52B11.2c 56 4x10�91 Yes

GDP-D-mannose pyrophosphorylasea O22287 D-mannose-1-P ? GDP-D-mannose C42C1.5c 55 3x10�143 Yes

GDP-D-mannose 3’,5’-epimerasea Q93VR3 GDP-D-mannose ? GDP-L-galactose D2096.4b 27 2x10�23 21stGDP-L-galactose phosphorylasea Q8RWE8 GDP-L-galactose ? L-galactose-1-P C10F3.4/MCP-1a 25 3x10�20 Yes

L-Galactose-1-P phosphatasea Q9M8S8 L-galactose-1-P ? L-galactose F13G3.5a 40 7x10�55 Yes

L-Galactose dehydrogenasea O81884 L-galactose ? L-galactono-1,4-lactone F37C12.12d 37 3x10�65 Yes

L-Galactono-1,4-lactonedehydrogenasea

Q9SU56 L-galactono-1,4-lactone ? L-ascorbate Y50D7A.7b 30 4x10�5 7th

Protein existence according to Uniprot.a Evidence at protein level.b Evidence at transcript level.c Inferred from homology.d Predicted.

A.N. Patananan et al. / Archives of Biochemistry and Biophysics 569 (2015) 32–44 39

these results open the possibility that ascorbate synthesis in C. ele-gans may occur through a novel pathway.

Previous research in Chlamydomonas found that oxidative stresscan increase ascorbate levels [25]. We therefore hypothesized thatoxidative stress may increase ascorbate accumulation in C. elegansand incubated mixed populations of N2 animals with OP50 E. coliand 50, 100, or 200 lM paraquat, a dipyridyl oxidative stress agent.After a 7 day incubation, we observed no significant difference byGC–MS analysis in the ascorbate:serine ratio of treated animalsin comparison to nontreated animals (Fig. 7A). We also tested ani-mals incubated in the presence of 0.5% and 1% ethanol and foundno increase in ascorbate concentration (Fig. 7B). Unsurprisingly,ascorbate levels declined sharply when 10% ethanol was added, a

concentration associated with poor survival of C. elegans [61], fur-ther suggesting that C. elegans may synthesize ascorbate (Fig. 7B).

C. elegans synthesizes ascorbate

To directly demonstrate that C. elegans synthesizes ascorbateand to elucidate a possible pathway, we labeled a mixed popula-tion of animals with either [12C6]- or [13C6]-glucose-labeled E. colion plates or in liquid media. We then analyzed the fragmentationpattern of ascorbate purified by GC–MS from extracts of theseworms. In the 12C-labeled animals from both plates and liquid cul-tures, we detected the 449 m/z fragment corresponding to the all12C-ascorbate species, as well as a 451 m/z species that could

Fig. 6. The addition of ascorbate precursors from known ascorbate biosynthetic pathways does not increase ascorbate levels in C. elegans. (A–C) C. elegans membrane,cytosolic, and intact worm fractions were incubated with a final concentration of 6.7 mM L-gulono-1,4-lactone for 2 h at room temperature and analyzed by HPLC asdescribed in the ‘‘Materials and methods.’’ Asterisks (*) denote the elution position of ascorbate. (D) L1 larvae obtained from 80 NGM plates were incubated in a finalconcentration of 1.7 mM D-galactose, L-gulono-1,4-lactone, D-galacturonic acid, D-glucose, or D-glucuronic acid for 2 h at 20 �C and 160 rpm, before the samples were lysedand analyzed by reversed-phase HPLC according to the ‘‘Materials and methods’’ section. The variability in ascorbate elution between panels A–C and D is due to a change inHPLC column.

40 A.N. Patananan et al. / Archives of Biochemistry and Biophysics 569 (2015) 32–44

represent the hydrogenated lactone species that proceeds ascor-bate in the plant and animal pathways (Fig. 8, compare to the L-ascorbate standard in Fig. 2A). For the corresponding material from13C-labeled animals, we detected not only the 449 and 451 ions,but also the 455 and 457 m/z ions, which would represent the fully13C-labeled species (Fig. 8). In the control animals grown on12C-E. coli, there were no 455 or 457 m/z ions detected (Fig. 8).Because the 451 and 457 m/z ions that possibly correspond tothe hydrogenated lactone precursor of ascorbate were detectedat the same GC–MS elution time as ascorbate, it was not possibleto definitely assign these ions to this possible precursor. Finally,we searched for molecules at other elution times in the GC–MSthat could be precursors in the biosynthesis pathway by comparingthe fragmentation patterns across GC fractions of the 12C- and13C-labeled samples. Although we found several known speciesthat incorporated the 13C-label (such as amino acids), we did notidentify any possible precursors to ascorbate.

We then analyzed each of the ascorbate GC–MS fragments toconfirm that ascorbate was 13C-labeled. C. elegans incubated with12C-E. coli had ascorbate fragment species whose mass-to-chargeratios closely matched that of the NIST mass spectral library data-base (Figs. 2A and 9). Because the 117, 147, 205, 259, 304, 332, and449 m/z species correspond to 2, 2, 2, 4, 3, 4, and 6 carbon speciesof ascorbate, respectively, the incorporation of heavy carbon in allpositions would result in fragments of 119, 149, 207, 263, 307, 336,and 455 m/z. With the exception of the 147 m/z fragment, fulllabeling was observed for the remaining species in animals incu-bated on plates and in liquid culture with 13C-labeled bacteria

(Fig. 9). Combined, these data support the conclusion that C. ele-gans is capable of synthesizing ascorbate. However, as describedabove, we found no evidence that this biosynthesis utilized eitherthe known plant or animal pathways.

Conclusions

Ascorbate and erythroascorbate are important reducing agentsfound in many vertebrates, plants, and fungi. To our knowledge,no other study has confirmed the biosynthesis of ascorbate in aninvertebrate organism. However, work on crustaceans [62] andinsects, such as the corn borer Diatraea grandiosella [63], the cornearworm Helicoverpa zea [64,65], the moth Plusia signata [66],and fruit fly D. melanogaster [39] suggest that ascorbate is presentand has biological roles potentially associated with metamorpho-sis, aging, and preventing oxidative stress. Furthermore, a studyinvolving the parasitic nematode Brugia malayi found ascorbateto be important for the L3 to L4 molt [67]. In C. elegans, ascorbateis an important molecule associated with handling oxidative stress,although its role in influencing longevity in wild type animals isunclear [68–72]. The addition of ascorbate not only alleviatesdefective phenotypes in animals exposed to pro-oxidants [73],but can also increase survival and restore normal longevity to shortlived mutant animals with elevated reactive oxygen species (ROS)[74–76]. Interestingly, ascorbate abolishes life span extension inlong lived mutant animals with elevated superoxide and normalROS levels [77]. In the present study, ascorbate was confirmed to

Fig. 7. Paraquat and ethanol stress do not increase ascorbate levels in C. elegans. (A)A mixed population of animals was obtained from approximately six NGM platesand incubated in S-medium supplemented with OP50 E. coli and 0, 50, 100, and200 lM paraquat for 7 d as described in the ‘‘Materials and methods.’’ GC–MSanalysis was performed as described in the ‘‘Materials and methods’’ and theamount of ascorbate in samples was analyzed by comparing the EIC area ofascorbate (332 m/z) to that of serine (204 m/z). The points represent four indepen-dent samples, the horizontal line is the average, and the error bars denote standarddeviations. Statistical significance was determined using an unpaired, two-tailedStudent’s t-test. (B) Mixed populations of animals from 80 NGM plates wereincubated in a microcentrifuge tube at 20 �C and 160 rpm in M9 mediumsupplemented with 0, 1%, 5%, or 10% ethanol. After incubation for 39 h, the animalswere pelleted, washed, lysed, and analyzed by reversed-phase HPLC according tothe ‘‘Materials and methods’’ section. The data points represent two independentsamples and the horizontal line denotes the average. Statistical significance wasdetermined using an unpaired, two-tailed Student’s t-test.

A.N. Patananan et al. / Archives of Biochemistry and Biophysics 569 (2015) 32–44 41

exist in C. elegans by two independent methods, and the biosynthe-sis of ascorbate in this organism is supported by in vivo labelingexperiments. Therefore, this study supports the hypothesis thatinvertebrate organisms are also able to synthesize ascorbate.

The presence of ascorbate in C. elegans is perhaps unsurprisingbecause many of its proteins have mammalian homologs thatrequire ascorbate for proper function. In particular, the ascor-

bate-dependent prolyl 4-hydroxylase is found in C. elegans and isimportant for the hydroxylation of collagen in other animals[78,79]. Approximately 1% of the genes in C. elegans are devotedto the synthesis of collagen, which composes 80% of the cuticleexoskeleton and is replaced in each molting cycle [80–82]. Variousmutations in the dpy-18 and phy-2 genes that encode the prolyl 4-hydrolase result in abnormal morphology and lethality [78,79].Additionally, the gene let-268 encodes the only lysyl hydroxylasein C. elegans and is important for type IV collagen processing[83]. It is therefore reasonable to expect C. elegans to have ascor-bate in order to maintain these Fe2+/a-ketoglutarate-dependentdioxygenases and ensure the proper formation of the eggshelland cuticle. Finally, the mammalian enzymes associated with car-nitine, tyrosine, and peptide hormone metabolism that use ascor-bate as a cofactor, including e-N-trimethyl-L-lysine/c-butyrobetaine hydroxylases, dopamine beta hydroxylase, peptidyl-glycine alpha-amidating monooxygenase, and the 4-hydrox-yphenylpyruvate dioxygenase [17], are highly similar to theGBH-1/GBH-2, TBH-1, PGAL-1, and HPD-1 proteins, respectively,in C. elegans.

Other than serving as an antioxidant and protein cofactor,ascorbate may also have a symbiotic role in supporting the micro-bial environment of C. elegans. It has been suggested that C. elegans,like other organisms, can not only benefit from their microbiota interms of the nutrients, vitamins, and energy they provide, but theworm can also support bacteria with shelter and nutrients [84].Although it is thought that bacteria are unable to synthesize ascor-bate, E. coli has an ascorbate transporter that may enable the fer-mentation of ascorbate and its use as a carbon source underanaerobic conditions [85–88]. Furthermore, C. elegans has homo-logs to the nucleobase ascorbate transporter (NAT) family of pro-teins found in other organisms [89]. In a preliminary analysis ofused liquid M9 medium after a C. elegans culture, we were ableto find a concentration of approximately 1 lM ascorbate (datanot shown). Although this may be due to the release of ascorbatefrom the lysed cells of dead animals, it is also possible that C. ele-gans may actively export ascorbate, either to neutralize oxidativestresses in the environment or provide a carbon source to bacteria,organisms in which it is integrally entwined with in nature.

In our study, we did not observe ascorbate in the unused culturemedia alone or in the E. coli used to grow the animals. Only C. ele-gans extracts were found to contain ascorbate at average concen-trations ranging between 21 lM in mixed populations to 144 lMin egg preparations. Interestingly, preliminary experiments indi-cate ascorbate levels in mixed animal populations are not sig-nificantly different when they are fed heat killed OP50 bacteria(data not shown), indicating ascorbate synthesis may not dependon bacterial viability. However, it is possible that feeding C. elegansa bacterial strain other than OP50 may result in an alteration inascorbate levels. Previous studies have found that the bacterial dietof C. elegans can influence gene expression, fat storage, metabo-lism, and longevity [75,90–94], and future studies could use dietto dissect the ascorbate biosynthesis pathway. It is also interestingto note that the concentration of ascorbate drops from the highestin eggs to the lowest level in the L1 larva and mixed populations.This result suggests ascorbate is required at different levels duringdevelopment for either added protection against oxidative stress orfor specific enzymatic processes. The hypothesis that ascorbatelevels are linked with enzymatic activities is possibly supportedby the observation that the expression of one prolyl 4-hydroxylasecatalytic subunit, the PHY-3 isoform, is limited to the early embryo,late larval, and adult nematode stages [95]. Additionally, a recentstudy has determined that the expression of several genes associ-ated with collagen formation in C. elegans is regulated [96]. Moreresearch is needed to determine how ascorbate levels are regulatedin C. elegans.

Fig. 8. Incorporation of 13C into C. elegans ascorbate from 13C-labeled E. coli food. (A) Expected mass change in ascorbate and the possible reduced precursor species due touniform 13C labeling. (B) A mixed population of animals were incubated with 12C6-D-glucose or 13C6-D-glucose-labeled E. coli on plates or in liquid culture at 20 �C for 85 h asdescribed in the ‘‘Materials and methods’’ section and analyzed by GC–MS. The asterisk denotes the position of ascorbate.

42 A.N. Patananan et al. / Archives of Biochemistry and Biophysics 569 (2015) 32–44

Fig. 9. Ascorbate fragmentation pattern from C. elegans fed unlabeled or 13C-labeledE. coli. Mixed populations of animals were incubated with unlabeled OP50 E. coli orOP50 E. coli labeled with 13C6-D-glucose on (A) NGM plates or (B) in liquid culturefor 85 h at 20 �C and analyzed by GC–MS as described in Fig. 8 and the ‘‘Materialsand methods.’’ Ascorbate fragmentation data for animals incubated with unlabeledor 13C-labeled OP50 E. coli is shown in blue and red, respectively, and overlaid fordirect comparison. The peaks labeled a, b, c, d, e, f, and g correspond to majorascorbate fragmentation peaks shown in Fig. 2 and include the 117, 147, 205, 259,304, 332, and 449 m/z ions. A magnified view of these fragmentation peaks isshown in the inset figures in panels A and B, with the y-axis representing relativelyintensity and the x-axis denoting the m/z. (For interpretation of the references tocolor in this figure legend, the reader is referred to the web version of this article.)

A.N. Patananan et al. / Archives of Biochemistry and Biophysics 569 (2015) 32–44 43

Through the use of 13C-labeled E. coli, the present work has pro-vided evidence for the first time of ascorbate biosynthesis in aninvertebrate organism. Although 13C-labeled ascorbate wasobserved when C. elegans were grown on plates with 13C-labeledE. coli, unlabeled species were also present. Interestingly, 13C-in-corporation appeared to be greater for animals grown in liquid cul-ture in comparison to those grown on plates. Unpublished work byReinke et al.2 found only 1% of genes differentially expressed whencomparing animals grown under these two different conditions.Interestingly, the genes upregulated in liquid culture wereassociated with oxidative stress protection and included thioredoxin,glutathione-S-transferase, and the mitochondrial NADP transhydro-genase. It is therefore possible that genes linked to ascorbate synthe-sis may also be altered in liquid culture.

Finally, this study did not observe any significant alteration inthe ascorbate levels of C. elegans exposed to oxidative stresses.

2 Personal communication from Valerie Reinke, Frederick G. Mann, Christina MWhittle, and Jason D. Lieb. Data is available at the NCBI Gene Expression Omnibus(GEO) database under GSE21526.

.

Although these oxidative agents may be increasing ascorbatelevels, we may not be able to observe any concentration changedue to the use of ascorbate to neutralize these molecules. We alsofound that C. elegans incubated with precursors from the plant andanimal pathways also do not synthesize increased levels of ascor-bate. Because our study only incubated L1 larvae with these pre-cursors, an increased level of ascorbate may be observed usingdifferent larval stages or specific growth conditions. Nevertheless,a similar negative result was observed when D. melanogaster, anorganism that has ascorbate in its extracts, was incubated with L-gulono-1,4-lactone [39]. It is therefore possible that invertebratesshare a common ascorbate biosynthetic pathway that is sig-nificantly different from that found in animals, plants, and fungi.

Acknowledgments

This work was supported by a Ruth L. Kirschstein NationalResearch Service Award GM007185 to ANP, and by NIH grantGM026020, a Senior Scholar in Aging Award from the Ellison Med-ical Foundation, and funds from the Elizabeth and Thomas PlottChair in Gerontology of the UCLA Longevity Center to SGC. Wewould like to acknowledge Gregory Khitrov from the UCLA Mole-cular Instrumentation Center, Kym Faull, and Christopher Ryanfor their help and advice on GC-MS analysis. We would also liketo thank Brian Young and Carole Linster for their contributions tothe early stages of this work, and Alison Frand for helpfuldiscussions.

References

[1] J. Du, J.J. Cullen, G.R. Buettner, Biochim. Biophys. Acta 1826 (2012) 443–457.[2] G.R. Buettner, B.A. Jurkiewicz, Radiat. Res. 145 (1996) 532–541.[3] C.L. Linster, S.G. Clarke, Trends Plant Sci. 13 (2008) 567–573.[4] Q. Gao, L. Zhang, J. Plant Physiol. 165 (2008) 138–148.[5] C. Barth, M. De Tullio, P.L. Conklin, J. Exp. Bot. 57 (2006) 1657–1665.[6] P. Muller-Moule, T. Golan, K.K. Niyogi, Plant Physiol. 134 (2004) 1163–1172.[7] C. Pignocchi, C.H. Foyer, Curr. Opin. Plant Biol. 6 (2003) 379–389.[8] C. Barth, W. Moeder, D.F. Klessig, P.L. Conklin, Plant Physiol. 134 (2004) 1784–

1792.[9] C.G. Fraga, P.A. Motchnik, M.K. Shigenaga, H.J. Helbock, R.A. Jacob, B.N. Ames,

Proc. Natl. Acad. Sci. U.S.A. 88 (1991) 11003–11006.[10] S.J. Duthie, A. Ma, M.A. Ross, A.R. Collins, Cancer Res. 56 (1996) 1291–1295.[11] E.S. Wintergerst, S. Maggini, D.H. Hornig, Ann. Nutr. Metab. 50 (2006) 85–94.[12] A.C. Carr, B. Frei, Am. J. Clin. Nutr. 69 (1999) 1086–1107.[13] B.N. Ames, M.K. Shigenaga, T.M. Hagen, Proc. Natl. Acad. Sci. U.S.A. 90 (1993)

7915–7922.[14] G.P. McGregor, H.K. Biesalski, Curr. Opin. Clin. Nutr. Metab. Care 9 (2006)

697–703.[15] C. Goodyear-Bruch, J.D. Pierce, Am. J. Crit. Care 11 (2002) 543–551.[16] A. Ozer, R.K. Bruick, Nat. Chem. Biol. 3 (2007) 144–153.[17] S. Englard, S. Seifter, Annu. Rev. Nutr. 6 (1986) 365–406.[18] W.A. Dunn, G. Rettura, E. Seifter, S. Englard, J. Biol. Chem. 259 (1984) 10764–

10770.[19] C.J. Rebouche, Am. J. Clin. Nutr. 54 (1991) 1147S–1152S.[20] K. Hirota, G.L. Semenza, Biochem. Biophys. Res. Commun. 338 (2005) 610–

616.[21] N. Smirnoff, P.L. Conklin, F.A. Loewus, Annu. Rev. Plant Physiol. Plant Mol.

Biol. 52 (2001) 437–467.[22] N. Gest, H. Gautier, R. Stevens, J. Exp. Biol. 64 (2013) 33–53.[23] P. Streb, S. Aubert, E. Gout, R. Bligny, J. Exp. Bot. 54 (2003) 405–418.[24] C. Paciolla, F. Tommasi, Biol. Plant. 47 (2003) 387–393.[25] E.I. Urzica, L.N. Adler, M.D. Page, C.L. Linster, M.A. Arbing, D. Casero, M.

Pellegrini, S.S. Merchant, S.G. Clarke, J. Biol. Chem. 287 (2012) 14234–14245.[26] E. Tel-Or, M.E. Huflejt, L. Packer, Arch. Biochem. Biophys. 246 (1986) 396–

402.[27] C. Obinger, G. Regelsberger, A. Pircher, G. Strasser, G.A. Peschek, Physiol.

Plant. 104 (1998) 693–698.[28] A. Lorence, B.I. Chevone, P. Mendes, C.L. Nessler, Plant Physiol. 134 (2004)

1200–1205.[29] W.Y. Zhang, A. Lorence, H.A. Gruszewski, B.I. Chevone, C.L. Nessler, Plant

Physiol. 150 (2009) 942–950.[30] C.G. Bartoli, J.P. Yu, F. Gomez, L. Fernandez, L. McIntosh, C.H. Foyer, J. Exp. Bot.

57 (2006) 1621–1631.[31] K. Dabrowski, Experientia 50 (1994) 745–748.[32] R. Moreau, K. Dabrowski, J. Fish Biol. 57 (2000) 733–745.[33] I.B. Chatterjee, Science 182 (1973) 1271–1272.[34] M.E. Rice, E.J.K. Lee, Y.Y. Choy, J. Neurochem. 64 (1995) 1790–1799.

44 A.N. Patananan et al. / Archives of Biochemistry and Biophysics 569 (2015) 32–44

[35] G. Drouin, J.R. Godin, B. Page, Curr. Genomics 12 (2011) 371–378.[36] R.S. Carr, J.M. Neff, Anal. Chem. 52 (1980) 2428–2430.[37] R.S. Carr, M.B. Bally, P. Thomas, J.M. Neff, Anal. Chem. 55 (1983) 1229–1232.[38] K. Dabrowski, S. Hinterleitner, Analyst 114 (1989) 83–87.[39] H.R. Massie, M.E. Shumway, S.J. Whitney, S.M. Sternick, V.R. Aiello, Exp.

Gerontol. 26 (1991) 487–494.[40] S.D. Gupta, C.R. Chaudhuri, I.B. Chatterjee, Arch. Biochem. Biophys. 152

(1972) 889–890.[41] S.Z.H. Wong, B.Y. Ching, Y.R. Chng, W.P. Wong, S.F. Chew, Y.K. Ip, Ascorbic acid

biosynthesis and brackish water acclimation in the euryhaline freshwaterwhite-rimmed stingray, Himantura signifer, PLoS One 8 (2013).

[42] C.L. Linster, T.A. Gomez, K.C. Christensen, L.N. Adler, B.D. Young, C. Brenner,S.G. Clarke, J. Biol. Chem. 282 (2007) 18879–18885.

[43] C.L. Linster, E. Van Schaftingen, FEBS J. 274 (2007) 1–22.[44] M.Y. Lachapelle, G. Drouin, Genetica 139 (2011) 199–207.[45] A.A. Badejo, K. Wada, Y. Gao, T. Maruta, Y. Sawa, S. Shigeoka, T. Ishikawa, J.

Exp. Bot. 63 (2012) 229–239.[46] F. Agius, R. Gonzalez-Lamothe, J.L. Caballero, J. Munoz-Blanco, M.A. Botella, V.

Valpuesta, Nat. Biotechnol. 21 (2003) 177–181.[47] N. Smirnoff, Nat. Biotechnol. 21 (2003) 134–136.[48] A. Baroja-Mazo, P. del Valle, J. Rua, S. de Cima, F. Busto, D. de Arriaga, N.

Smirnoff, Fungal Genet. Biol. 42 (2005) 390–402.[49] V.A. Dumbrava, M.L. Pall, Biochim. Biophys. Acta 926 (1987) 331–338.[50] J.A. Nick, C.T. Leung, F.A. Loewus, Plant Sci. 46 (1986) 181–187.[51] W.K. Huh, B.H. Lee, S.T. Kim, Y.R. Kim, G.E. Rhie, Y.W. Baek, C.S. Hwang, J.S.

Lee, S.O. Kang, Mol. Microbiol. 30 (1998) 895–903.[52] V. Valpuesta, M.A. Botella, Trends Plant Sci. 9 (2004) 573–577.[53] L.N. Adler, T.A. Gomez, S.G. Clarke, C.L. Linster, J. Biol. Chem. 286 (2011)

21511–21523.[54] L.N. Adler, From vitamin C to metabolite repair: the role of novel sugar

nucleotide phosphorylases, Ph.D. Thesis, Biochemistry and Molecular Biology,UCLA, 2012.

[55] M.F. Oberbacher, H.M. Vines, Nature 197 (1963) 1203–1204.[56] C.R. Dawson, K.G. Strothkamp, K.G. Krul, Ann. N. Y. Acad. Sci. 258 (1975) 209–

220.[57] C.L. Linster, L.N. Adler, K. Webb, K.C. Christensen, C. Brenner, S.G. Clarke, J.

Biol. Chem. 283 (2008) 18483–18492.[58] P.L. Conklin, S.A. Saracco, S.R. Norris, R.L. Last, Genetics 154 (2000) 847–856.[59] R. Sasaki, T. Kurokawa, S. Tero-Kubota, J. Gerontol. 38 (1983) 26–30.[60] J. Lykkesfeldt, T.M. Hagen, V. Vinarsky, B.N. Ames, FASEB J. 12 (1998) 1183–

1189.[61] R. Dhawan, D.B. Dusenbery, P.L. Williams, J. Toxicol. Environ. Health A 58

(1999) 451–462.[62] G. Merchie, P. Lavens, P. Sorgeloos, Aquaculture 155 (1997) 165–181.[63] G.M. Chippendale, J. Nutr. 105 (1975) 499–507.[64] G.W. Felton, C.B. Summers, Arch. Insect Biochem. Physiol. 29 (1995) 187–197.[65] C.B. Summers, G.W. Felton, Biochim. Biophys. Acta 1156 (1993) 235–238.[66] P.N. Saxena, Nature 209 (1966) 535–536.[67] T.V. Rajan, N. Paciorkowski, I. Kalajzic, C. McGuiness, J. Parasitol. 89 (2003)

868–870.

[68] K. Pallauf, J.K. Bendall, C. Scheiermann, K. Watschinger, J. Hoffmann, T.Roeder, G. Rimbach, Food Chem. Toxicol. 58 (2013) 255–263.

[69] T.J. Schulz, K. Zarse, A. Voigt, N. Urban, M. Birringer, M. Ristow, Cell Metab. 6(2007) 280–293.

[70] A. Shibamura, T. Ikeda, Y. Nishikawa, Mech. Ageing Dev. 130 (2009) 652–655.[71] J. Gruber, L.F. Ng, S.K. Poovathingal, B. Halliwell, FEBS Lett. 583 (2009) 3377–

3387.[72] L.A. Harrington, C.B. Harley, Mech. Ageing Dev. 43 (1988) 71–78.[73] S. Khare, T. Gomez, C.L. Linster, S.G. Clarke, Mech. Ageing Dev. 130 (2009)

670–680.[74] J.Z. Huang, B.D. Lemire, J. Mol. Biol. 387 (2009) 559–569.[75] S.S. Pang, S.P. Curran, Cell Metab. 19 (2014) 221–231.[76] S.S. Pang, D.A. Lynn, J.Y. Lo, J. Paek, S.P. Curran, Nat. Commun. 5 (2014) 5048.[77] W. Yang, S. Hekimi, PLoS Biol. 8 (2010) e1000556.[78] L. Friedman, J.J. Higgin, G. Moulder, R. Barstead, R.T. Raines, J. Kimble, Proc.

Natl. Acad. Sci. U.S.A. 97 (2000) 4736–4741.[79] A.D. Winter, A.P. Page, Mol. Cell. Biol. 20 (2000) 4084–4093.[80] A.R. Frand, S. Russel, G. Ruvkun, PLoS Biol. 3 (2005) 1719–1733.[81] The C. elegans Sequencing Consortium, Science 283 (1999) 35.[82] G.N. Cox, M. Kusch, R.S. Edgar, J. Cell Biol. 90 (1981) 7–17.[83] K.R. Norman, D.G. Moerman, Dev. Biol. 227 (2000) 690–705.[84] F. Cabreiro, D. Gems, EMBO Mol. Med. 5 (2013) 1300–1310.[85] W.B. Esselen, J.E. Fuller, J. Bacteriol. 37 (1939) 501–521.[86] R.M. Young, L.H. James, J. Bacteriol. 44 (1942) 75–84.[87] H.E. Richter, J. Switala, P.C. Loewen, Can. J. Microbiol. 34 (1988) 822–825.[88] Z. Zhang, M. Aboulwafa, M.H. Smith, M.H. Saier Jr., J. Bacteriol. 185 (2003)

2243–2250.[89] C. Gournas, I. Papageorgiou, G. Diallinas, Mol. BioSyst. 4 (2008) 404–416.[90] D.A. Garsin, J.M. Villanueva, J. Begun, D.H. Kim, C.D. Sifri, S.B. Calderwood, G.

Ruvkun, F.M. Ausubel, Science 300 (2003) 1921.[91] P.L. Larsen, C.F. Clarke, Science 295 (2002) 120–123.[92] F. Gomez, G.C. Monsalve, V. Tse, R. Saiki, E. Weng, L. Lee, C. Srinivasan, A.R.

Frand, C.F. Clarke, BMC Microbiol. 12 (2012).[93] K.K. Brooks, B. Liang, J.L. Watts, PLoS One 4 (2009).[94] L.S. Yilmaz, A.J.M. Walhout, Trends Genet. 30 (2014) 496–503.[95] P. Riihimaa, R. Nissi, A.P. Page, A.D. Winter, K. Keskiaho, K.I. Kivirikko, J.

Myllyharju, J. Biol. Chem. 277 (2002) 18238–18243.[96] G.J. Hendriks, D. Gaidatzis, F. Aeschimann, H. Grosshans, Mol. Cell 53 (2014)

380–392.[97] S.K. Olson, G. Greenan, A. Desai, T. Muller-Reichert, K. Oegema, J. Cell Biol. 198

(2012) 731–748.[98] K.C. Justi, J.V. Visentainer, N.E. de Souza, M. Matsushita, Arch. Latinoam. Nutr.

50 (2000) 405–408.[99] S. Tulipani, B. Mezzetti, F. Capocasa, S. Bompadre, J. Beekwilder, C.H.R. De Vos,

E. Capanoglu, A. Bovy, M. Battino, J. Agric. Food Chem. 56 (2008) 696–704.[100] S.Y. Wang, K.S. Lewers, L. Bowman, M. Ding, J. Am. Soc. Hortic. Sci. 132 (2007)

647–658.[101] C.S. Tsao, P.Y. Leung, M. Young, J. Nutr. 117 (1987) 291–297.[102] C.M. Spickett, N. Smirnoff, A.R. Pitt, Free Radic. Biol. Med. 28 (2000) 183–192.